Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. A method implemented on at least one device each of which has at least one processor and a storage device, the method comprising: obtaining, by the at least one processor, a first start point and a first end point relating to a road network, wherein the road network includes multiple links, and the first start point and the first end point are acquired from a location device; obtaining, by the at least one processor, a route recommendation model, wherein a parameter of the route recommendation model includes weights of the multiple links of the road network, the weights are determined based at least partially on a model different from the route recommendation model, and the model for determining the weights is updated based on adjustment values of the weights, wherein the route recommendation model is generated based on a method for generating a recommended route, the method comprising: obtaining a second start point and a second end point relating to the road network; obtaining a plurality of historical routes from a storage device, each of the plurality of historical routes including the second start point and the second end point; determining a plurality of links between the second start point and the second end point; determining primary weights of the plurality of links; determining one or more ordinary routes from the second start point to the second end point based, at least in part, on the primary weights of the plurality of links; adjusting primary weights of links on the one or more ordinary routes; adjusting primary weights of links on the historical routes; and determining the route recommendation model based on the adjusted weights of the links on the one or more ordinary routes and the adjusted weights of the links on the historical routes; and determining, by the at least one processor, a recommended route from the first start point to the first end point based on the route recommendation model; and causing a user terminal to display the recommended route on an user interface.
This invention relates to a navigation system that generates optimized routes for users based on historical route data and dynamic link weighting. The system addresses the challenge of providing accurate and efficient route recommendations by leveraging machine learning models to adjust the importance of road segments (links) in a network. The method involves obtaining a user's start and end points from a location device and using a pre-trained route recommendation model. This model incorporates link weights that are initially determined by a separate model, which updates these weights based on historical route data and user behavior. The route recommendation model is generated by analyzing historical routes between similar start and end points, calculating primary weights for the links involved, and then refining these weights based on both ordinary routes (computed without historical data) and the historical routes themselves. The adjusted weights are then used to determine the optimal route for the user's current start and end points. The final recommended route is displayed on a user terminal. This approach improves route accuracy by dynamically adjusting link weights based on real-world usage patterns, ensuring more reliable navigation suggestions.
2. The method of claim 1 , wherein determining the primary weights of the plurality of links includes: obtaining road network data related to the road network, the road network data including one or more road features of the plurality of links between the second start point and the second end point; and determining the primary weights of the plurality of links based on the one or more road features of the plurality of links.
This invention relates to a method for determining primary weights of links in a road network to optimize routing or navigation. The method addresses the challenge of accurately assessing the importance or priority of different road segments in a network to improve pathfinding algorithms. The primary weights are used to influence routing decisions, ensuring that the most relevant or efficient paths are selected based on real-world road conditions. The method involves obtaining road network data that includes various road features of the links between a specified start and end point. These road features may include attributes such as road type (e.g., highway, local road), speed limits, traffic conditions, historical travel times, or other relevant characteristics. By analyzing these features, the method calculates primary weights for each link, which quantify their significance in the network. These weights can then be used in routing algorithms to prioritize certain paths over others, improving navigation efficiency and accuracy. The approach ensures that routing decisions are based on up-to-date and contextually relevant road data, enhancing the reliability of navigation systems.
3. The method of claim 1 , wherein the one or more ordinary routes are determined according to maximum margin planning (MMP) algorithm or maximum entropy inverse reinforcement learning (IRL) algorithm.
This invention relates to route planning for autonomous vehicles or navigation systems, specifically improving route selection by incorporating advanced algorithms to determine optimal paths. The core problem addressed is the need for more efficient and adaptive route planning that considers dynamic factors like traffic, road conditions, and driver preferences beyond traditional shortest-path or fastest-path methods. The method involves determining one or more ordinary routes for a vehicle using either a maximum margin planning (MMP) algorithm or a maximum entropy inverse reinforcement learning (IRL) algorithm. The MMP algorithm optimizes routes by maximizing the margin between the chosen path and alternative routes, ensuring robustness against uncertainties. The maximum entropy IRL algorithm learns preferences or objectives from observed behavior, allowing the system to infer and incorporate implicit user or system priorities into route selection. These algorithms enable the system to adapt to varying conditions and user preferences dynamically, improving efficiency and user satisfaction. The invention enhances traditional route planning by leveraging machine learning and optimization techniques to make routes more reliable and personalized. This approach is particularly useful in autonomous driving, logistics, and navigation applications where adaptability and efficiency are critical.
4. The method of claim 1 , wherein the one or more ordinary routes are determined according to the MMP algorithm, the one or more ordinary routes being the shortest route from the second start point to the second end point.
This invention relates to route determination in network systems, specifically addressing the challenge of efficiently identifying optimal paths between network nodes. The method involves determining one or more ordinary routes between a second start point and a second end point, where these routes are calculated as the shortest paths using the MMP (Minimum Message Path) algorithm. The MMP algorithm is a pathfinding technique that evaluates network topology to identify the most efficient routes based on minimizing message transmission time or distance. The ordinary routes are derived by applying the MMP algorithm to the network graph, ensuring that the selected paths are the shortest possible between the specified endpoints. This approach optimizes network performance by reducing latency and resource consumption in data transmission. The method may be used in various network applications, including routing protocols, traffic management, and distributed systems, where efficient path selection is critical. By leveraging the MMP algorithm, the invention provides a systematic way to determine optimal routes, enhancing overall network efficiency and reliability.
5. The method of claim 1 , wherein adjusting the primary weights of the links on the one or more ordinary routes and the primary weights on the historical routes includes: increasing the primary weights of the links on the historical routes; and decreasing the primary weights of the links on the one or more ordinary routes.
Network routing systems dynamically adjust link weights to optimize traffic flow, but existing methods often fail to account for historical traffic patterns, leading to suboptimal routing decisions. This invention improves routing efficiency by dynamically adjusting primary weights of links on both ordinary and historical routes. The method increases the primary weights of links on historical routes, prioritizing paths that have been frequently used in the past, while decreasing the primary weights of links on ordinary routes to deprioritize less frequently used paths. This adjustment ensures that traffic is routed more efficiently by leveraging past usage data, reducing congestion and improving network performance. The system may also include mechanisms to detect and handle network failures, ensuring reliable routing even under adverse conditions. By dynamically balancing weights based on historical and current traffic patterns, the invention enhances overall network efficiency and reliability.
6. The method of claim 1 , further including: obtaining a loss function associated with the route recommendation model; determining whether a convergence of the loss function is reached; in response to the determination that the convergence of the loss function is reached, outputting the route recommendation model to the terminal device; and in response to the determination that the convergence of the loss function is not reached, initating a next iteration to update the route recommendation model, wherein the next iteration includes: determining another one or more ordinary routes from the second start point to the second end point based, at least in part, on the primary weights of the plurality of links; adjusting primary weights of links on the another one or ore ordinary routes; adjusting the primary weights of links on the historical routes; and determining the route recommendation model based on the adjusted primary weights of the links on the another one or more ordinary routes and the adjusted primary weights of the links on the historical routes.
The invention relates to optimizing route recommendation models for navigation systems. The problem addressed is improving the accuracy and efficiency of route recommendations by dynamically adjusting link weights based on historical and ordinary routes. The method involves training a route recommendation model by iteratively refining link weights until a convergence criterion is met. During each iteration, the system determines one or more ordinary routes from a start point to an end point using current primary weights of links. It then adjusts the primary weights of links on these ordinary routes and also adjusts weights of links on historical routes. The model is updated based on these adjusted weights. This process repeats until the loss function associated with the model converges, at which point the finalized model is output to a terminal device. The approach ensures that the model learns from both historical data and dynamically generated routes, improving recommendation accuracy over time. The system avoids overfitting by continuously refining weights until optimal performance is achieved.
7. The method of claim 1 , further including: obtaining a preset number of iterations; determining whether the preset number of iterations are performed; in response to the determination that the preset number of iterations are performed, outputting the route recommendation model to the terminal device; and in response to the determination that the preset number of iterations are not performed, initating a next iteration to update the route recommendation model, wherein the next iteration includes; determining new one or more ordinary routes from the second start point to the second end point based, at least in part, on the primary weights of the plurality of links; adjusting primary weights of links on the new one or more ordinary routes; adjusting the primary weights of links on the historical routes; and determining the route recommendation model based on the adjusted primary weights of the links on the new one or more ordinary routes and the adjusted primary weights of the links on the historical routes.
This invention relates to route recommendation systems, specifically improving the accuracy and efficiency of route models by iteratively refining link weights based on historical and newly determined routes. The problem addressed is the static nature of traditional route recommendation models, which fail to adapt to dynamic traffic conditions or user preferences over time. The method involves generating a route recommendation model by iteratively updating link weights in a transportation network. Initially, a preset number of iterations is defined. For each iteration, the system determines one or more ordinary routes between a start and end point based on current primary weights assigned to network links. These weights are then adjusted for links on both the newly determined routes and historical routes. The route recommendation model is updated based on these adjusted weights. If the preset iteration count is reached, the final model is output to a terminal device. Otherwise, the process repeats with another iteration, refining the model further. This iterative approach ensures the model continuously improves by incorporating new route data while retaining valuable historical information, leading to more accurate and adaptive route recommendations.
8. The method of claim 1 , wherein the weights of the multiple links of the road network relate to a parameter of the route recommendation model.
A system and method for optimizing route recommendations in a road network involves analyzing multiple links within the network to improve navigation accuracy. The method assigns weights to these links based on specific parameters of a route recommendation model, which may include factors such as traffic conditions, road quality, or historical travel data. By adjusting these weights, the system dynamically refines route suggestions to enhance efficiency and reliability. The approach ensures that the model adapts to real-time changes, providing users with the most optimal paths. This technique is particularly useful in navigation applications where dynamic adjustments are necessary to account for varying road conditions and user preferences. The method may also incorporate additional data sources, such as weather or road construction updates, to further refine the weighting process. The overall goal is to improve the accuracy and adaptability of route recommendations in complex road networks.
9. The method of claim 1 , wherein the weights of the multiple links of the road network may be determined based on at least one of a road width, a road length, a traffic flow, traffic lights, road safety, intersections, the direction of the road, or a road type.
This invention relates to optimizing navigation or routing within a road network by dynamically determining the weights of multiple links in the network. The problem addressed is the need for more accurate and efficient route calculations that account for various real-world factors affecting travel time and safety. Traditional routing algorithms often rely on static or limited data, leading to suboptimal paths. The method involves assigning weights to road network links based on multiple factors to improve routing decisions. These factors include road width, length, traffic flow, traffic lights, road safety, intersections, road direction, and road type. By incorporating these variables, the system can generate more realistic and adaptive routes. For example, a narrow road with high traffic flow may receive a higher weight, indicating it should be avoided when possible, while a wider road with fewer intersections may be prioritized. Traffic lights and road safety metrics further refine the weighting to account for delays and hazard levels. The direction of the road and its type (e.g., highway, urban street) also influence the weight, ensuring the routing algorithm considers bidirectional constraints and specialized road characteristics. This approach enhances navigation systems by providing more context-aware and efficient routing solutions.
10. A system, comprising: at least one storage medium including a set of instructions; and at least one processor configured to communicate with the at least one storage medium, wherein when executing the set of instructions, the at least one processor is directed to: obtain a first start point and a first end point relating to a road network, wherein the road network includes multiple links, and the first start point and the first end point are acquired from a location device; obtain a route recommendation model, wherein a parameter of the route recommendation model includes weights of the multiple links of the road network, the weights are determined based at least partially on a model different from the route recommendation model, and the model for determining the weights is updated based on adjustment values of the weights, wherein to obtain the route recommendation model, the at least one processor is directed to: obtain a second start point and a second end point relating to the road network: obtain a plurality of historical routes from a storage device, each of the plurality of historical routes including the second start point and the second end point; determine a plurality of inks between the second start point and the second end point; determine primary weights of the plurality of links; determine one or more ordinary routes from the second start point to the second end point based, at least in part, on the primary weights of the plurality of links: adjust primary weights of links on the one or more ordinary routes; adjust primary weights of links on the historical routes; and determine the route recommendation model based on the adjusted weights of the links on the one or more ordinary routes and the adjusted weights of the links on the historical routes; and determine a recommended route from the first start point to the first end point based on the route recommendation model; and cause a user terminal to display the recommended route on a user interface.
The system provides a route recommendation solution for road networks by leveraging historical route data and dynamic weight adjustments. The technology addresses the challenge of optimizing route recommendations by incorporating real-world usage patterns and dynamically adjusting link weights in the road network. The system obtains a start and end point from a location device and uses a route recommendation model that assigns weights to road network links. These weights are determined by a separate model that updates based on adjustment values. To build the route recommendation model, the system collects historical routes between a second set of start and end points, identifies all possible links between them, and assigns primary weights to these links. It then determines ordinary routes based on these weights and adjusts the weights of links on both ordinary and historical routes. The adjusted weights are used to finalize the route recommendation model. For a given start and end point, the system calculates a recommended route using this model and displays it on a user terminal. The dynamic adjustment of link weights based on historical and ordinary routes improves the accuracy and relevance of route recommendations.
11. The system of claim 10 , wherein to determine the primary weights of the plurality of links, the at least one processor is directed to: obtain road network data related to the road network, the road network data including one or more road features of the plurality of links between the second start point and the second end point; and determine the primary weights of the plurality of links based on the one or more road features of the plurality of links.
A system for route optimization in a road network analyzes road features to determine primary weights for links between a start and end point. The system obtains road network data, which includes various characteristics of the links, such as road type, traffic conditions, or infrastructure details. These features are used to calculate primary weights, which represent the relative importance or influence of each link in route planning. The system then applies these weights to optimize routing decisions, ensuring efficient and accurate pathfinding. The primary weights may be adjusted dynamically based on real-time or historical data to improve routing accuracy. This approach enhances navigation systems by incorporating detailed road network attributes into route calculations, leading to more reliable and adaptive travel recommendations. The system is particularly useful in applications requiring precise route optimization, such as logistics, autonomous vehicles, or real-time navigation services. By leveraging road features, the system provides a more nuanced and context-aware routing solution compared to traditional methods that rely solely on distance or time metrics.
12. The system of claim 10 , wherein the one or more ordinary routes are determined according to the MMP algorithm, the one or more ordinary routes being the shortest route from the second start point to the second end point.
This invention relates to route optimization systems, specifically for determining efficient paths in a network or transportation system. The problem addressed is the need to identify optimal routes that balance computational efficiency with path quality, particularly when multiple routes must be evaluated. The system includes a route optimization module that calculates one or more ordinary routes between a second start point and a second end point. These routes are determined using the MMP (Modified Multi-Path) algorithm, which ensures the shortest path is selected. The MMP algorithm likely involves evaluating multiple potential paths while prioritizing the shortest route based on predefined criteria such as distance, time, or resource constraints. The system may also include a primary route optimization module that calculates a primary route between a first start point and a first end point, which could involve different optimization techniques or constraints. The ordinary routes may be used as fallback or alternative paths when the primary route is unavailable or suboptimal. The invention is particularly useful in applications where multiple route options must be considered, such as logistics, navigation, or network routing, where both efficiency and reliability are critical. The use of the MMP algorithm ensures that the shortest viable path is identified, improving overall system performance.
13. The system of claim 10 , wherein to adjust the primary weights of the links on the one or more ordinary routes and the primary weights on the historical routes, the at least one processor is directed to: increase the primary weights of the links on the historical routes; and decrease the primary weights of the links on the one or more ordinary routes.
Network routing systems dynamically adjust link weights to optimize traffic flow, but existing methods often fail to account for historical route performance, leading to suboptimal path selection. This invention improves routing efficiency by dynamically adjusting primary weights of links on both ordinary and historical routes. The system identifies historical routes that have demonstrated reliable performance and increases their link weights to prioritize them in future routing decisions. Simultaneously, it decreases the primary weights of links on ordinary routes to deprioritize them, ensuring that traffic is directed toward more reliable paths. This adjustment process is performed by at least one processor, which evaluates route performance data to determine which links should be weighted higher or lower. The system thereby enhances network reliability and reduces congestion by leveraging historical route performance to inform real-time routing decisions. This approach is particularly useful in large-scale networks where route selection impacts overall system efficiency.
14. The system of claim 10 , the at least one processor is further directed to: obtain a loss function associated with the route recommendation model; determine whether a convergence of the loss function is reached; in response to the determination that the convergence of the loss function is reached, output the route recommendation model to the terminal device; and in response to the determination that the convergence of the loss function is not reached, initate a next iteration to update the route recommendation model, wherein the next iteration includes: determine another one or more ordinary routes from the second start point to the second end point based, at least in part, on the primary weights of the plurality of links; adjust primary weights of links on the another one or more ordinary routes; adjust the primary weights of links on the historical routes; and determine the route recommendation model based on the adjusted primary weights of the links on the another one or more ordinary routes and the adjusted primary weights of the links on the historical routes.
A system for optimizing route recommendations in navigation applications addresses the challenge of providing accurate and efficient route suggestions based on dynamic traffic conditions and user preferences. The system includes a route recommendation model that processes historical route data and real-time traffic information to generate optimal routes between a start point and an end point. The model assigns primary weights to links (road segments) in a transportation network, where these weights influence route selection by reflecting factors such as travel time, congestion, or user preferences. During operation, the system obtains a loss function associated with the route recommendation model to evaluate its performance. If the loss function converges, indicating the model has achieved sufficient accuracy, the system outputs the model to a terminal device (e.g., a user's navigation app). If convergence is not reached, the system initiates another iteration to refine the model. In this iteration, the system determines additional ordinary routes (alternative paths) from the start to the end point based on the current primary weights of the links. It then adjusts the primary weights of links on these new routes and also updates the weights of links on previously identified historical routes. The model is then recalculated using these adjusted weights, improving its accuracy for future route recommendations. This iterative process continues until the model converges, ensuring the most efficient and reliable route suggestions.
15. The system of claim 10 , the at least one processor is further directed to: obtain a preset number of iterations; determine whether the preset number of iterations are performed; in response to the determination that the preset number of iterations are performed, output the route recommendation model to the terminal device; and in response to the determination that the preset number of iterations are not performed, initate a next iteration to update the route recommendation model, wherein the next iteration includes: determine new one or more ordinary routes from the second start point to the second end point based, at least in part, on the primary weights of the plurality of links; adjust primary weights of links on the new one or more ordinary routes; adjust the primary weights of links on the historical routes; and determine the route recommendation model based on the adjusted primary weights of the links on the new one or more ordinary routes and the adjusted primary weights of the links on the historical routes.
The invention relates to a route recommendation system that iteratively refines a route recommendation model to improve navigation suggestions. The system addresses the challenge of providing accurate and optimized route recommendations by dynamically adjusting link weights based on historical and newly determined routes. The system operates by first obtaining a preset number of iterations for model refinement. During each iteration, the system determines new routes from a start point to an end point using primary weights assigned to links in a network. These weights are adjusted based on the newly identified routes and historical routes. The system then updates the route recommendation model using the adjusted weights. If the preset number of iterations is completed, the refined model is output to a terminal device. If not, the system initiates the next iteration, repeating the process of route determination, weight adjustment, and model refinement. This iterative approach ensures the model continuously improves by incorporating both historical data and new route information, leading to more accurate and efficient route recommendations.
16. The method of claim 10 , wherein the weights of the multiple links of the road network relate to a parameter of the route recommendation model.
A system and method for optimizing route recommendations in a road network involves analyzing multiple links within the network to improve navigation accuracy and efficiency. The road network is represented as a graph where each link has associated weights that correspond to specific parameters of a route recommendation model. These weights influence the selection of optimal routes by the model, ensuring that recommended paths align with desired criteria such as travel time, distance, or traffic conditions. The method dynamically adjusts these weights based on real-time or historical data to enhance the model's performance. By incorporating these weighted links, the system provides more accurate and personalized route suggestions, improving user experience and navigation reliability. The approach leverages graph-based representations of road networks to model complex traffic scenarios and optimize routing decisions. This technique is particularly useful in applications requiring adaptive and intelligent navigation solutions, such as autonomous vehicles, logistics planning, and real-time traffic management systems. The method ensures that the route recommendation model remains responsive to changing conditions, enhancing its effectiveness in diverse environments.
17. The system of claim 10 , wherein the weights of the multiple links of the road network may be determined based on at least one of a road width, a road length, a traffic flow, traffic lights, road safety, intersections, the direction of the road, or a road type.
This invention relates to a system for analyzing and optimizing road networks, particularly for applications in navigation, traffic management, or autonomous vehicle routing. The system models a road network as a graph where roads are represented as links with assigned weights, which influence pathfinding and traffic flow analysis. The weights of these links are dynamically determined based on multiple factors, including road width, length, traffic flow patterns, the presence and timing of traffic lights, road safety metrics, intersection complexity, road directionality, and road type (e.g., highway, urban street). By incorporating these variables, the system provides a more accurate representation of real-world driving conditions, enabling better route optimization, congestion prediction, and traffic management. The weighted graph structure allows for efficient computation of optimal paths while accounting for physical and operational constraints of the road network. This approach improves upon traditional methods that rely solely on distance or time-based metrics, offering a more nuanced and adaptive solution for navigation and traffic analysis.
18. A non-transitory computer readable medium storing instructions, the instructions, when executed by a computing device, causing the computing device to: obtain a first start point and a first end point relating to a road network, wherein the road network includes multiple links, and the first start point and the first end point are acquired from a location device; obtain a route recommendation model, wherein a parameter of the route recommendation model includes weights of the multiple links of the road network, and the weights are determined based at least partially on a model different from the route recommendation model, and the model for determining the weights is updated based on adjustment values of the weights wherein to obtain the route recommendation model, the at least one processor is directed to: obtain a second start point and a second end point relating to the road network; obtain a plurality of historical routes from a storage device, each of the plurality of historical routes including the second start point and the second end point; determine a plurality of links between the second start point and the second end point; determine primary weights of the plurality of links; determine one or more ordinary routes from the second start point to the second end point based, at least in part, on the primary weights of the plurality of links; adjust primary weights of links on the one or more ordinary routes; adjust primary weights of links on the historical routes; and determine the route recommendation model based on the adjusted weights of the links on the one or more ordinary routes and the adjusted weights of the links on the historical routes; and determine a recommended route from the first start point to the first end point based on the route recommendation model; and cause a user terminal to display the recommended route on a user interface.
The invention relates to a system for generating and displaying recommended routes in a road network using a machine learning-based approach. The problem addressed is the need for accurate and efficient route recommendations that consider historical travel patterns and dynamically adjust link weights in the road network to improve routing accuracy. The system obtains a start and end point from a location device and uses a route recommendation model to determine the optimal path. The model incorporates weights for each link in the road network, which are initially determined by a separate model and then refined through an iterative process. To build the route recommendation model, the system analyzes historical routes between a second set of start and end points, identifies the links involved, and assigns primary weights to these links. It then generates ordinary routes based on these weights and adjusts the weights of links on both the ordinary routes and the historical routes. The final route recommendation model is derived from these adjusted weights. For a given user query, the system uses the refined model to calculate the best route from the first start to end point and displays it on a user interface. The dynamic adjustment of link weights based on historical data ensures that the recommendations adapt to real-world travel patterns, improving accuracy over time. This approach enhances navigation systems by leveraging machine learning to optimize routing decisions.
Unknown
September 22, 2020
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.